FIELD OF INVENTION
The present invention relates generally to a sensor system and method for operating such a system. More particularly, the invention relates to a sensor system having systematic and stochastic phase errors.
Meinke, Gundlach, Taschenbuch der Hochfrequenztechnik, 5.sup.th Edition, Springer Verlag, pp. S3-S4, describes an FMCW radar as distance or velocity sensor. The abbreviation FM stands for frequency modulation and CW stands for continuous wave. A signal source generates a frequency-modulated signal that propagates wave-shaped. The signal source includes, a microwave oscillator and a modulator. The preferably linear frequency-modulated signal is beamed out by an antenna and reflected a subject under test. The received signal is mixed in a mixer with the transmission signal present at the moment. The difference frequency that derives is a criterion for the distance of the subject under test from the antenna. A monostatic or a bistatic antenna arrangement can be utilized for transmitting and receiving the frequency modulated signal In. Given the monostatic arrangement, the transmission signal and the reception signal are beamed out or, received by the common transmission and reception antenna. The transmission signal is separated from the mixer with a circulator or directional coupler and the reception signal separated from the signal source is conducted to the mixer. In the bistatic antenna arrangement, separate transmission and reception antennas are provided. A sensor system in which a high-precision frequency modulation is generated with a control loop that has a delay element is disclosed by German Document No. 27 10 841 A1.
FIG. 3 shows a known sensor system with a delay line for generating a reference signal. The signal source MO is shown supplying a signal that propagates wave-shaped and that is frequency-modulated. The signal is preferably linearly modulated. The transmission signal s(t) is supplied to a transmission and reception diplexer SEW that, for example, can be a circulator or a directional coupler. From this diplexer, the transmission signal proceeds into the transmission and reception unit SEE that has one or more antennas available to it for emitting and for receiving the microwave signal. After the transmission signal has been emitted and reflected from a target, it is supplied from the transmission and reception means to a mixer means EMIX as reception signal r(t) via the transmission and reception diplexer, the reception signal being mixed with the transmission signal s(t) in the mixer to form the measured signal mess(t). Subsequently, disturbing high-frequency mix components are eliminated from the measured signal, preferably in a low-pass filter TP.
The transmission signal s(t) coming from the signal source MO is delayed with a delay means V. The signal delayed by .tau..sub.ref is mixed with the transmission signal s(t) in a further mixer means REFMIX to form the reference signal ref(t) that is then preferably conducted through a low-pass filter. The measured signal mess(t) and the reference signal ref(t) are supplied to an evaluation means AE.
FIG. 4 shows the corresponding sensor system to FIG. 3 with a bistatic antenna arrangement. In this system the transmission and reception diplexer is eliminated. The transmission and reception arrangement is two separate antennas for transmission and reception.
German Document No. 195 33 124 discloses an apparatus for distance measurement with a signal source that comprises a modulator and a voltage-controlled oscillator. The oscillator generates a frequency-modulated signal. Generally, the modulation is not ideally linear. The transmission signal s(t) is emitted by the antenna, reflected at the target and received. In a first mixer, the signal r(t).about.s(t-.tau..sub.mess) received after the time .tau..sub.mess is mixed with the current transmission signal.
This branch of the apparatus serving as measuring means, referred to below as MES, has a reference means, referred to below as REF, coordinated with it. This REF is supplied with the frequency-modulated signal generated by the signal source. The REF contains a surface wave component that forwards the frequency-modulated signal to a second mixer delayed by the time duration .tau..sub.ref. This mixer generates the reference signal ref(t). The low-pass filtered signals mess (t) and ref(t) are supplied to an evaluation unit.
As already presented, the fundamental FMCW principle is present when a linearly frequency-modulated signal s(t) is emitted and the echo signals r(t) reflected by the target and received are mixed back with the transmission signal s(t). T is the duration of the sweep event with t.epsilon.[0,T]. Due to the time delay .tau..sub.mess that the echo signals exhibit compared to the transmission signal, a constant frequency .sub.Figure proportional to the distance or, a linearly rising phase swing arises as mix product given a linear sweep. Given a non-linear sweep, clear deviations from these ideal conditions are present.
It is assumed in the consideration of a non-linear sweep that the linear sweep with the basic radian frequency .omega..sub.o and a sweep rate .mu. of ##EQU1##
deviates from an ideal phase linearity with a phase error .DELTA..phi., i.e.: ##EQU2##
Leaving higher mix products and constant phase amounts out of consideration, the mix product of s(t) and r(t) thus derives: EQU mess(t)=cos [.mu..multidot..tau..multidot.t+.DELTA..phi.(t+.tau.-.DELTA..phi.(t)].
It is assumed in the consideration of the error that the phase errors in the interval .tau. can be assumed as linear changes. Under this pre-condition, the ideal signal frequency ##EQU3##
is respectively distorted at time t by the noise term ##EQU4##
linearized by t. Given a non-linear sweep, the signal frequency is no longer constant but distributed over a broad frequency range. The noise effects rise proportionally with the signal running time and, thus, proportionally with the measurement distance.
When the time-dependent noise term .DELTA..PHI.(t,.tau.).sub.ref for an arbitrary reference distance s.sub.ref (with the appertaining signal running time .tau..sub.ref) is known, for example from a reference measurement, then the phase errors for arbitrary measurement distances s.sub.mess (with the appertaining signal running time s.sub.mess) can also be derived therefrom according to ##EQU5##
It follows from the preceding presentation that the momentary phase .phi.(t) of the signal is proportional to the measurement distance or, respectively, to the signal running time. Thus: ##EQU6##
German Document No. 195 33 124 and the publication of Vossiek et al., "Novel FMCW radar system concept with adaptive compensation of phase errors", 26.sup.th European Microwave Conference, Prague, Czech Republic, Sep., 9-12 1996, pp. 135-139, disclose various methods of how a distorted signal can be distortion-corrected such given a known phase curve that the phase errors are corrected.
One possibility of signal equalization is comprised in sampling the measured signal not in constant time intervals, as usual, but in constant phase intervals (preferably zero-axis crossings, i.e. phases spaced at 180.degree. relative to one another). The measured signal is thereby sampled at times t.sub.n at which the phase of the reference signal .phi..sub.ref has changed by a multiple of a predetermined phase step width n.multidot.d.phi..sub.ref, i.e. EQU mess'(n .multidot.dt')=mess(t.sub.n =t.vertline..sub..phi..sub..sub.ref .sub.(t)=n.multidot.d.phi..sub..sub.ref ).
In order to determine the new, corrected measured signal mess'(n .multidot.dt'), the sampling points acquired in this way are sequentially arranged in the time sampling interval dt'. The sampling interval dt' is calculated from d.phi..sub.ref, s.sub.ref and the sweep rate .mu.. This linearization or, equalization method is called post-equalization below.
Another linearization method includes calculating a corrected modulation signal on the basis of the phase errors measured in the reference unit. This linearization or, equalization method is called pre-equalization below. Phase errors in the following sweeps can be corrected by driving the oscillator with the modified control signal, however only those phase errors that occur systematically, i.e. temporally reproducible. Stochastic phase errors (principally the phase noise of the oscillator) are not corrected by this method. However, the stochastic phase errors are often highly disturbing, especially given signals with a long running time (i.e. given distance sensors with targets at a great distance).
When target and sensor move relative to one another, then the distance-dependent frequency f.sub.dist of the measured signal has an additional Doppler frequency f.sub.dop superimposed on it. The signal spectrum of mess (t) is thus shifted by the Doppler frequency f.sub.dop in this case. The Doppler frequency is calculated according to: ##EQU7##
whereby v represents the subject velocity and .lambda. represents the wavelength of the measured signal.
Since the reference path no longer correctly describes the real measurement path, the signal of the post-equalization method no longer ensues ideally in this case. Errors mainly occur given short signal running times since the distance-dependent frequencies are relatively small compared to the Doppler frequency. When the Doppler frequency is high compared to distance-dependent frequencies, the error in the signal equalization due to the Doppler shift is negligible or, it is assured that the proposed signal equalization leads to an improvement of the measured result.